<-40 



A IN AIR, WITH CONTAMINATED EXTERIOR 

 D COVERED, WITH CLEAN EXTERIOR 



O IMMERSED IN TAP WATER 

 X GLASS CITRATE BOTTLE 



Figure 37. — Change of salinity and weight with time. 



When analyzing the first and second bottles of group 2, they were 

 weighed first, then washed. Just the converse was done for the 

 final bottle; it was washed first, then weighed. It is believed that 

 some of the weight gain was caused by its immersion in water (although 

 toweled dry before weighing), but probably this was not enough to 

 explain the entire increase. Those in group 3 oscillated between 

 gains and losses in both weight and salinity, but they tended more 

 toward a salinity decrease with a corresponding weight increase. 

 It is of course recognized that polytliylene bottles difi^er in construction 

 and thus that each may vary from another to a certain extent in its 

 permeability. When the salinities of the citrate bottles were 

 measured, the second bottle's salinity was very close to the original 

 value, whereas the third bottle's salinity was higher by 0.01 °/oo. 



The batch of Copenhagen water used to standardize the WHOI 

 salinity bridge was changed between ^ = 61 and ^ = 83 days; therefore, 

 because of this change there might have existed a small relative differ- 

 ence between the readings prior to and subsequent to this. This 

 might explain the small difference between the last citrate reading 

 and the first two. An equally plausible explanation of the salinity 

 change in the citrate bottle could be evaporation occurring around the 

 spring-operated lid. 



For each sample the change of weight divided by the initial weight 

 was plotted (fig. 38) against the change of salinity and then the groups 

 were compared. Lines connecting the origin with the average of the 



75 



